© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 43991
    Prev
    Next
    INDEX
    Explanations

    occurrences of the "th" digraph

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    er
    -0.15
    hte
    -0.15
    ER
    -0.15
    erah
    -0.15
    coni
    -0.15
    ez
    -0.15
    anja
    -0.14
    Æ°á»
    -0.14
    asaki
    -0.14
    otton
    -0.14
    POSITIVE LOGITS
    ttp
    0.26
    entication
    0.22
    orne
    0.22
    ompson
    0.21
    à¥įà¤ł
    0.21
    istle
    0.20
    letics
    0.20
    aniel
    0.20
    uis
    0.19
    o
    0.19
    Activations Density 0.030%

    No Known Activations